Publications-Theses

Article View/Open

Publication Export

Google ScholarTM

NCCU Library

Citation Infomation

Related Publications in TAIR

題名 普適提演算法比較
作者 陳炳霖
貢獻者 余清祥<br>張源俊
<br>
陳炳霖
關鍵詞 普適提
日期 2003
上傳時間 2009-09-14
摘要 普適提(Boosting)是近年來發展迅速且廣泛被應用於分類問題的方法之一,
其特點是利用精確度較低或是較粗糙的分類方法(weak learner)為基礎,
經由多次反覆分類之後,
將其結果合併而提升分類的精確度。
本研究為探討若是SVM, KNN, LDA等分類基底時,
普適提是否仍然能夠達到改進的效果,
或是會造成模型過適的情況(overfitting)。
我們發現到普適提在訓練資料集(training set)與測試資料集(testing set)的解釋變數
獨立且同分佈(independent identically ditributed)產生的情況下,
無論用任何一種分類法則皆有造成模型過適的可能;
但是,若訓練資料集與測試資料集的相關程度很高的情況下,
則不會發生模型過適的情況,
因此,欲探討不同相關係數以及分布類似的程度對測試資料集的結果有何影響。
參考文獻 Dettling, M. and Buhlmann, P. (2002) How to use boosting for tumor
classi&#64257;cation with gene expression data.
Escudero, G. , Marquez, L. and Rigau, G. (2000) Boosting applied to
word sense disambiguation. In LNAI 1810: Proceedings of the 12th
European Conference on Machine Learning, ECML, pages 129-141.
Freund, Y. (2001) An adaptive version of the boost by majority algorithm.
Machine Learning, 43(3):293-318.
Freund, Y. and Schapire, R.E. (1999) A short introduction to boosting.
Journal of Japanese Society for Artificial Intelligence, 14(5):771-
780.
Hastie, T., Tibshirani, R. and Friedman, J. (2001) The Elements of
Statistical Learning: data mining, inference and prediction.
Lebanon, G. and La&#64256;erty, J. (2001) Boosting and maximum likelihood
for exponential models. In Neural Information Processing Systems
(NIPS), volume 14.
Long, P.M. (2002) Minimum majority classi&#64257;cation and boosting. In
AAAI
Lugosi, G. and Vayatis, N. (2002) A consistent strategy for boosting
algorithms. In Proceedings of the Annual Conference on Computational
Learning Theory, volume 2375 of LNAI, pages 303-318.
Mannor, S. and Meir, R. (2001) Weak learners and improved convergence
rate in boosting. In Advances in Neural Information Processing
Systems 13: Proc.NIPS.
Mannor, S., Meir, R. and Mendelson, S. (2001) On the consistency of
boosting algorithms. submitted to Advances in Neural Information
Processing 14.
Meir, R. and Ratsch, G. (2003) An introduction to boosting and leveraging.
In S. Mendelson and A. Smola, editors, Advanced Lectures
on Machine Learning, LNCS, pages 119-184.
Onoda, T., Ratsch, G. and Muller, K.-R. (2000) Applying support
vector machines and boosting to a non-intrusive monitoring system
for household electric appliances with inverters.
Ratsch, G., Mika, S. , Scholkopf, B. and Muller, K.-R. (2000) Constructing
boosting algorithms from SVMs: an application to oneclass
classi&#64257;cation. IEEE PAMI, 24(9).
Ratsch, G., Scholkopf, B. , Mika, S. and Muller, K.-R. (2000) SVM
and Boosting: One class. Technical Report 119, GMD FIRST,
Berlin.
Ratsch, G. and Warmuth, M.K. (2002) Maximizing the margin with
boosting. In Proceedings of the Annual Conference on Computational
Learning Theory, volume 2375 of LNAI, pages 334-350.
Ratsch, G. andWarmuth, M.W. (2002) E&#64259;cient margin maximization
with boosting.
Schapire, R.E. (1999) A brief introduction to boosting. In Proceedings
of the Sixteenth International Joint Conference on Artificial
Intelligence.
Yaniv, R.E., Meir, R. and David, S.B. (2000) Localized boosting.
In Proceedings of the 13th Annual Conference on Computational
Learning Theory, pages 190-199.
描述 碩士
國立政治大學
統計研究所
91354025
92
資料來源 http://thesis.lib.nccu.edu.tw/record/#G0091354025
資料類型 thesis
dc.contributor.advisor 余清祥<br>張源俊zh_TW
dc.contributor.advisor <br>en_US
dc.contributor.author (Authors) 陳炳霖zh_TW
dc.creator (作者) 陳炳霖zh_TW
dc.date (日期) 2003en_US
dc.date.accessioned 2009-09-14-
dc.date.available 2009-09-14-
dc.date.issued (上傳時間) 2009-09-14-
dc.identifier (Other Identifiers) G0091354025en_US
dc.identifier.uri (URI) https://nccur.lib.nccu.edu.tw/handle/140.119/30890-
dc.description (描述) 碩士zh_TW
dc.description (描述) 國立政治大學zh_TW
dc.description (描述) 統計研究所zh_TW
dc.description (描述) 91354025zh_TW
dc.description (描述) 92zh_TW
dc.description.abstract (摘要) 普適提(Boosting)是近年來發展迅速且廣泛被應用於分類問題的方法之一,
其特點是利用精確度較低或是較粗糙的分類方法(weak learner)為基礎,
經由多次反覆分類之後,
將其結果合併而提升分類的精確度。
本研究為探討若是SVM, KNN, LDA等分類基底時,
普適提是否仍然能夠達到改進的效果,
或是會造成模型過適的情況(overfitting)。
我們發現到普適提在訓練資料集(training set)與測試資料集(testing set)的解釋變數
獨立且同分佈(independent identically ditributed)產生的情況下,
無論用任何一種分類法則皆有造成模型過適的可能;
但是,若訓練資料集與測試資料集的相關程度很高的情況下,
則不會發生模型過適的情況,
因此,欲探討不同相關係數以及分布類似的程度對測試資料集的結果有何影響。
zh_TW
dc.description.tableofcontents 1 緒論
1.1 簡介
1.2 研究目標與方法
2 普適提演算法介紹
2.1 普是提演算法流程
2.1.1 AdaBoost
2.1.2 LogitBoost
2.1.3 其它Boosting
2.2 classifier的簡介
2.2.1 SVM
2.2.2 KNN
2.2.3 LDA
3 模擬研究與結果分析
3.1 研究方法與設定
3.2 Independent identically distributed data
3.2.1 Master Card Type
3.2.2 同心圓
3.2.3 棋盤格
3.2.4 Micky Mouse Type
3.2.5 Symmetric
3.3 Correlated data
3.3.1 High Correlation
3.3.2 Small Error
3.3.3 相關係數與誤差標準差的選擇
3.4 小結
4 結論與建議
4.1 本研究之結論
4.2 後續研究之建議
參考文獻
附錄
zh_TW
dc.language.iso en_US-
dc.source.uri (資料來源) http://thesis.lib.nccu.edu.tw/record/#G0091354025en_US
dc.subject (關鍵詞) 普適提zh_TW
dc.title (題名) 普適提演算法比較zh_TW
dc.type (資料類型) thesisen
dc.relation.reference (參考文獻) Dettling, M. and Buhlmann, P. (2002) How to use boosting for tumorzh_TW
dc.relation.reference (參考文獻) classi&#64257;cation with gene expression data.zh_TW
dc.relation.reference (參考文獻) Escudero, G. , Marquez, L. and Rigau, G. (2000) Boosting applied tozh_TW
dc.relation.reference (參考文獻) word sense disambiguation. In LNAI 1810: Proceedings of the 12thzh_TW
dc.relation.reference (參考文獻) European Conference on Machine Learning, ECML, pages 129-141.zh_TW
dc.relation.reference (參考文獻) Freund, Y. (2001) An adaptive version of the boost by majority algorithm.zh_TW
dc.relation.reference (參考文獻) Machine Learning, 43(3):293-318.zh_TW
dc.relation.reference (參考文獻) Freund, Y. and Schapire, R.E. (1999) A short introduction to boosting.zh_TW
dc.relation.reference (參考文獻) Journal of Japanese Society for Artificial Intelligence, 14(5):771-zh_TW
dc.relation.reference (參考文獻) 780.zh_TW
dc.relation.reference (參考文獻) Hastie, T., Tibshirani, R. and Friedman, J. (2001) The Elements ofzh_TW
dc.relation.reference (參考文獻) Statistical Learning: data mining, inference and prediction.zh_TW
dc.relation.reference (參考文獻) Lebanon, G. and La&#64256;erty, J. (2001) Boosting and maximum likelihoodzh_TW
dc.relation.reference (參考文獻) for exponential models. In Neural Information Processing Systemszh_TW
dc.relation.reference (參考文獻) (NIPS), volume 14.zh_TW
dc.relation.reference (參考文獻) Long, P.M. (2002) Minimum majority classi&#64257;cation and boosting. Inzh_TW
dc.relation.reference (參考文獻) AAAIzh_TW
dc.relation.reference (參考文獻) Lugosi, G. and Vayatis, N. (2002) A consistent strategy for boostingzh_TW
dc.relation.reference (參考文獻) algorithms. In Proceedings of the Annual Conference on Computationalzh_TW
dc.relation.reference (參考文獻) Learning Theory, volume 2375 of LNAI, pages 303-318.zh_TW
dc.relation.reference (參考文獻) Mannor, S. and Meir, R. (2001) Weak learners and improved convergencezh_TW
dc.relation.reference (參考文獻) rate in boosting. In Advances in Neural Information Processingzh_TW
dc.relation.reference (參考文獻) Systems 13: Proc.NIPS.zh_TW
dc.relation.reference (參考文獻) Mannor, S., Meir, R. and Mendelson, S. (2001) On the consistency ofzh_TW
dc.relation.reference (參考文獻) boosting algorithms. submitted to Advances in Neural Informationzh_TW
dc.relation.reference (參考文獻) Processing 14.zh_TW
dc.relation.reference (參考文獻) Meir, R. and Ratsch, G. (2003) An introduction to boosting and leveraging.zh_TW
dc.relation.reference (參考文獻) In S. Mendelson and A. Smola, editors, Advanced Lectureszh_TW
dc.relation.reference (參考文獻) on Machine Learning, LNCS, pages 119-184.zh_TW
dc.relation.reference (參考文獻) Onoda, T., Ratsch, G. and Muller, K.-R. (2000) Applying supportzh_TW
dc.relation.reference (參考文獻) vector machines and boosting to a non-intrusive monitoring systemzh_TW
dc.relation.reference (參考文獻) for household electric appliances with inverters.zh_TW
dc.relation.reference (參考文獻) Ratsch, G., Mika, S. , Scholkopf, B. and Muller, K.-R. (2000) Constructingzh_TW
dc.relation.reference (參考文獻) boosting algorithms from SVMs: an application to oneclasszh_TW
dc.relation.reference (參考文獻) classi&#64257;cation. IEEE PAMI, 24(9).zh_TW
dc.relation.reference (參考文獻) Ratsch, G., Scholkopf, B. , Mika, S. and Muller, K.-R. (2000) SVMzh_TW
dc.relation.reference (參考文獻) and Boosting: One class. Technical Report 119, GMD FIRST,zh_TW
dc.relation.reference (參考文獻) Berlin.zh_TW
dc.relation.reference (參考文獻) Ratsch, G. and Warmuth, M.K. (2002) Maximizing the margin withzh_TW
dc.relation.reference (參考文獻) boosting. In Proceedings of the Annual Conference on Computationalzh_TW
dc.relation.reference (參考文獻) Learning Theory, volume 2375 of LNAI, pages 334-350.zh_TW
dc.relation.reference (參考文獻) Ratsch, G. andWarmuth, M.W. (2002) E&#64259;cient margin maximizationzh_TW
dc.relation.reference (參考文獻) with boosting.zh_TW
dc.relation.reference (參考文獻) Schapire, R.E. (1999) A brief introduction to boosting. In Proceedingszh_TW
dc.relation.reference (參考文獻) of the Sixteenth International Joint Conference on Artificialzh_TW
dc.relation.reference (參考文獻) Intelligence.zh_TW
dc.relation.reference (參考文獻) Yaniv, R.E., Meir, R. and David, S.B. (2000) Localized boosting.zh_TW
dc.relation.reference (參考文獻) In Proceedings of the 13th Annual Conference on Computationalzh_TW
dc.relation.reference (參考文獻) Learning Theory, pages 190-199.zh_TW